Skip to content
Studio

Images and video,
made where the rest of your work lives.

Write a shot once and render it with 10 image and 10 video models — clips up to 30 seconds, 6 models with sound. Every take lands in your Library, ready for the next build.

Get early accessSee pricing
Studio

Three dancers in coral, cobalt and saffron silk. Morning light through tall windows.

Video16:91080p8 sGenerate
How a shot gets made

Five steps,
one panel.

  1. 01

    Write the shot once

    Describe it in plain words, or start from a preset that writes the prompt for you.

  2. 02

    Pick an image or video model

    10 image and 10 video models sit in the same panel.

  3. 03

    Only valid options

    The panel offers just what that model accepts — its shapes, resolutions, clip lengths and sound.

  4. 04

    Compare takes

    Ask for a variation and put it next to the last one before you keep it.

  5. 05

    Keep it, use it

    Results land in your Library with the model and settings that made them — ready for chat and Build.

Only valid options

The panel knows what each model takes.

Pick Veo 3.1 and you see its 2 shapes, its resolutions and its clip length — nothing it would reject. Switch model and the choices change with it.

Veo 3.1Google
Shape
16:99:16
Resolution
720p1080p4K
Length
4, 6 or 8 s
Sound
On
The roster

20 models,
every option they accept.

Filter by kind. Each card is the model's real setting list from the Studio catalogue.

  • GPT Image 2.5 SunburstOpenAI
    image
    Shapes
    8 + auto
    Resolution
    —
    1:13:22:34:33:416:99:1621:9
  • Nano Banana ProGoogle
    image
    Shapes
    10
    Resolution
    1K · 2K · 4K
    1:12:33:23:44:34:55:49:16+2
  • Nano Banana 2Google
    image
    Shapes
    14
    Resolution
    512 · 1K · 2K · 4K
    1:11:41:82:33:23:44:14:3+6
  • Seedream 5.0 ProByteDance Seed
    image
    Shapes
    17 + auto
    Resolution
    1K · 2K
    1:11:22:12:33:23:44:34:5+9
  • FLUX.2 MaxBlack Forest Labs
    image
    Shapes
    8 + auto
    Resolution
    —
    1:14:33:43:22:316:99:1621:9
  • Qwen Image 3 ProQwen
    image
    Shapes
    13
    Resolution
    1K · 2K
    1:11:21:42:12:33:23:44:1+5
  • MAI-Image-2.6Microsoft
    image
    Shapes
    7 + auto
    Resolution
    —
    1:14:33:416:99:163:22:3
  • Krea 2 LargeKrea
    image
    Shapes
    7
    Resolution
    1K
    1:14:33:216:94:52:39:16
  • Grok Imagine Image 2.0xAI
    image
    Shapes
    13 + auto
    Resolution
    1K · 2K
    1:13:44:39:1616:92:33:29:19.5+5
  • Recraft V4.1 ProRecraft
    image
    Shapes
    5 + auto
    Resolution
    —
    1:14:33:416:99:16
  • Seedance 2.5ByteDance
    video
    Shapes
    6
    Resolution
    480p · 720p
    Length
    4–30 s
    Sound
    Yes
    16:94:31:13:49:1621:9
  • Veo 3.1Google
    video
    Shapes
    2
    Resolution
    720p · 1080p · 4K
    Length
    4, 6 or 8 s
    Sound
    Yes
    16:99:16
  • Kling 3.0 ProKuaishou
    video
    Shapes
    3
    Resolution
    720p
    Length
    3–15 s
    Sound
    Yes
    16:99:161:1
  • Sora 2 ProOpenAI
    video
    Shapes
    2
    Resolution
    720p · 1080p
    Length
    4, 8, 12, 16 or 20 s
    Sound
    Yes
    16:99:16
  • Runway Gen-4.5Runway
    video
    Shapes
    2
    Resolution
    720p
    Length
    2–10 s
    Sound
    —
    16:99:16
  • Wan 3.0 PrimeAlibaba
    video
    Shapes
    5
    Resolution
    480p · 720p · 1080p
    Length
    2–30 s
    Sound
    Yes
    16:94:31:13:49:16
  • Hailuo 3 MaxMiniMax
    video
    Shapes
    6
    Resolution
    768p · 480p
    Length
    5–15 s
    Sound
    —
    21:916:94:31:13:49:16
  • FLUX.3 VideoBlack Forest Labs
    video
    Shapes
    6
    Resolution
    720p · 1080p
    Length
    5–20 s
    Sound
    Yes
    21:916:94:31:13:49:16
  • Grok Imagine Video 1.5xAI
    video
    Shapes
    7
    Resolution
    480p · 720p · 1080p
    Length
    1–15 s
    Sound
    —
    16:99:161:14:33:43:22:3
  • HappyHorse 1.1Alibaba
    video
    Shapes
    7
    Resolution
    720p · 1080p
    Length
    3–15 s
    Sound
    —
    16:99:161:14:33:421:99:21

Generation runs through your OpenRouter account. Live pricing is checked before each generation, and availability is checked with the provider each time.

Presets

Start from a format,
not a blank prompt.

Twenty video presets from the app — pick one, add your product, and Studio writes the prompt.

Everyday product reviewUGC · 9:16
Sports gear reviewUGC · 9:16
Phone accessory demoTikTok · 9:16
Sneaker unboxingUGC · 9:16
Everyday creator storyInstagram · 9:16
Problem to solutionTikTok · 9:16
Beauty routine hookUGC · 9:16
Unbox and try onInstagram · 9:16
Hands-on ASMRUGC · 9:16
Packaging detailsBrand · 9:16
Sneaker try-onInstagram · 9:16
At-home lifestyleUGC · 9:16
Selfie testimonialUGC · 9:16
Direct-to-camera adBrand · 9:16
Quick how-to hookTikTok · 9:16
Cinematic product launchBrand · 9:16
Food brand showcaseBrand · 9:16
Lifestyle campaignBrand · 9:16
Before and afterTikTok · 9:16
Skincare tutorialInstagram · 9:16
Made in Studio

Real outputs,
playing as they came out.

Silk in motion
An image made in Studio: Still · composition
Still · composition
Fractured core
An image made in Studio: Still · variation
Still · variation
Camera study
An image made in Studio: Still · product
Still · product
Slow push-in
An image made in Studio: Still · first version
Still · first version
Light study
An image made in Studio: Still · light
Still · light
Films

Short films,
in our own style.

From our launch film: painted keyframes, animated with Seedance 2.5 — one of Studio's video models.

The treeRoots of light under a city at night.
The libraryA wall of jars, each one holding a memory.
The forgeSparks on the anvil, gears turning.
The constellationA hand drawing lines between stars.
The threadOne line of light through the dark.
The ideaThe moment a plan clicks.

Not another tab.
A studio inside the app you build in.

  • Twenty models, one panel

    10 image and 10 video models — switch without leaving the shot.

  • Twenty presets

    UGC, TikTok, Instagram and brand formats that write the prompt for you.

  • Options that can't be wrong

    Each model only shows the settings it really accepts.

  • A Library, not a downloads folder

    Every result keeps its model and settings, next to your code and games.

  • Straight into the work

    Hand a clip or image to the agent in chat, or use it in the next build.

  • Cut the film here

    Order clips on a timeline, trim each scene, export MP4 · H.264 · 30 fps.

Questions

Which models can I use?
GPT Image 2.5 Sunburst, Nano Banana Pro, Nano Banana 2, Seedream 5.0 Pro, FLUX.2 Max, Qwen Image 3 Pro, MAI-Image-2.6, Krea 2 Large, Grok Imagine Image 2.0, Recraft V4.1 Pro for images; Seedance 2.5, Veo 3.1, Kling 3.0 Pro, Sora 2 Pro, Runway Gen-4.5, Wan 3.0 Prime, Hailuo 3 Max, FLUX.3 Video, Grok Imagine Video 1.5, HappyHorse 1.1 for video. All models
How is generation paid for?
Generation runs through your OpenRouter account. The app checks the live price before each generation, so you see it before you spend it.
Which video models make sound?
Seedance 2.5, Veo 3.1, Kling 3.0 Pro, Sora 2 Pro, Wan 3.0 Prime, FLUX.3 Video.
How long can a clip be?
It depends on the model — the longest offers 30 seconds. The panel only shows the lengths each model accepts.
Can I cut a finished film?
Yes. Put clips in order on the timeline, trim each scene's in and out points, and export an MP4 (H.264, 30 fps).
Where do my images and videos go?
Into your Library, with the model and settings that made them. Save one, or use it in chat to hand it to the agent.

Write the shot once.
Render it with every model.

Get early accessSee pricing